This application addresses Dissemination and Implementation Research in Health (R21); Exploratory/Developmental Research (PAR-10-040) Health systems want to know which implementation strategies will have the greatest population impact so they can make wise use of limited funds. This application proposes to identify the key bottlenecks that are barriers to the implementation of recommended diabetes and dyslipidemia screening for adults with mental disorders in order to identify the best intervention strategies. Among the most economically-disadvantaged with serious mental illness, excess cardiovascular risk contributes to decades of reduced life expectancy. Physicians prescribing second-generation antipsychotics (SGA) will be studied to focus the analysis because SGAs are commonly taken by adu]lts with mental disorders, they increase metabolic risk themselves, specific monitoring is recommended and data show severe under-screening in clinical practice. The study tests four physician practice domains hypothesized to affect screening rates: knowledge; attitudes; physical-behavioral health care coordination; and practice recall and reminder systems. We aim to identify implementation target groups (i.e.segments) based on the patient-mix, provider and practice factors most strongly associated with screening (Aim 2). This project focuses on the upstream issue of screening because it is where there is a big initial barrier to care, and screening is the gateway to treatment. However, downstream treatment barriers are also critical. We aim to estimate how often no follow-up care occurs among patients with abnormal lab values and if barriers affecting screening also affect the likelihood of follow-up (Aim 3). To reach these aims, a unique and comprehensive dataset of Medicaid providers and patient claims data from the State of Missouri will be analyzed. These data will be combined with a survey of physician knowledge, attitudes and practice characteristics and with electronic medical record data from community mental health centers. For Aims 1 and 2, the primary outcome is baseline and annual glucose and lipid testing using laboratory claims. Analysis for Aim 1 uses multivariate models designed to measure the extent that the factors describe variation in screening tests. Results of Aim 1 will inform the selection of segmentation factors used in the Aim 2 cluster analysis. Split-sample validation will be used to assess the stability of the segments. For Aim 3, the analytic approach for Aim 1 will be used for the outcome of follow-up care (defined as a repeat lab test, primary care office visit, or initiation o a cholesterol-lowering or oral diabetes drug). An Advisory Board of public health stakeholders and academic experts in mental health and implementation sciences will advise the research team in interpreting the findings and selecting the implementation strategies to move forward into testing. Missouri Medicaid is a unique research environment to achieve this objective because a disproportionate number of adults with mental disorders receive care through Medicaid, the research team has an active collaboration with Missouri stakeholders, and significant interest exists within Missouri to apply the research learning to conduct a randomized intervention trial to test implementation strategies.